You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by Hermes Flying <fl...@yahoo.com> on 2012/11/26 21:36:15 UTC

Tomcat ssl vulnerability CVE-2009-3555

Hi,
I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 

You notein your docs that this is not a Tomcat issue per se, but JSSE issue.
Please note that allowUnsafeLegacyRenegotiation is set to false. Looking into the source code I see the following:

public void handshake(Socket sock) throws IOException {

    ((SSLSocket)sock).startHandshake(); 

    if(!allowUnsafeLegacyRenegotiation) {
       // disable all ciphers, avoiding any subsequent handshake
       ((SSLSocket)sock).setEnabledCipherSuites(new String[0]);
    }
}

Also looking into Tomcat6/7 source code I see that the only difference is to check for existence of TLS_EMPTY_RENEGOTIATION_INFO_SCSV in JSSE. 

But other than that the logic is the same. So I can only assume that upgrading to Tomcat 6 would not solve my problem.
Since I have verified via debugging also that allowUnsafeLegacyRenegotiation is indeed false, why do I get reports on SSL client renegotiation vulnerability? 


I see in your notes (http://tomcat.apache.org/security-5.html) that:
Requires JRE that supports RFC 5746. For Oracle JRE that is known to be 6u22 or later. 


But as I understand the code, even if JRE does not support the RFC you still disable the handshake with your "hack" to set no enable ciphers


Is this a Java/JSSE problem? What can I do? Can you please help me?


Thank you!

Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by Hermes Flying <fl...@yahoo.com>.
Hi, yes I am sorry for not mentioning but I am using BIO connector. 


I am using IBM JRE 6. 

Not Oracle/Sun. That is why I am trying to figure out what component (Tomcat/Java) is the one that has the issue.
I am not 100% it is the latest.




________________________________
From: Daniel Mikusa <dm...@vmware.com>
To: Tomcat Users List <us...@tomcat.apache.org> 
Sent: Tuesday, November 27, 2012 12:09 AM
Subject: Re: Tomcat ssl vulnerability CVE-2009-3555

On Nov 26, 2012, at 3:36 PM, Hermes Flying wrote:

> Hi,
> I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 
> 
> You notein your docs that this is not a Tomcat issue per se, but JSSE issue.
> Please note that allowUnsafeLegacyRenegotiation is set to false.

Are you using the BIO connector?  This option is specific to that connector.  It would be helpful if you could share your connector configuration.

What version of the JVM are you using?  Can you upgrade to a version that resolves this issue?

Dan



> Looking into the source code I see the following:
> 
> public void handshake(Socket sock) throws IOException {
> 
>     ((SSLSocket)sock).startHandshake(); 
> 
>     if(!allowUnsafeLegacyRenegotiation) {
>        // disable all ciphers, avoiding any subsequent handshake
>        ((SSLSocket)sock).setEnabledCipherSuites(new String[0]);
>     }
> }
> 
> Also looking into Tomcat6/7 source code I see that the only difference is to check for existence of TLS_EMPTY_RENEGOTIATION_INFO_SCSV in JSSE. 
> 
> But other than that the logic is the same. So I can only assume that upgrading to Tomcat 6 would not solve my problem.
> Since I have verified via debugging also that allowUnsafeLegacyRenegotiation is indeed false, why do I get reports on SSL client renegotiation vulnerability? 
> 
> 
> I see in your notes (http://tomcat.apache.org/security-5.html) that:
> Requires JRE that supports RFC 5746. For Oracle JRE that is known to be 6u22 or later. 
> 
> 
> But as I understand the code, even if JRE does not support the RFC you still disable the handshake with your "hack" to set no enable ciphers
> 
> 
> Is this a Java/JSSE problem? What can I do? Can you please help me?
> 
> 
> Thank you!

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org

Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by Hermes Flying <fl...@yahoo.com>.
Hi, yes I am sorry for not mentioning but I am using BIO connector. 

I am using IBM JRE 6. 

Not Oracle/Sun. That is why I am trying to figure out what component (Tomcat/Java) is the one that has the issue.
I am not 100% it is the latest.




________________________________
 From: Daniel Mikusa <dm...@vmware.com>
To: Tomcat Users List <us...@tomcat.apache.org> 
Sent: Tuesday, November 27, 2012 12:09 AM
Subject: Re: Tomcat ssl vulnerability CVE-2009-3555
 
On Nov 26, 2012, at 3:36 PM, Hermes Flying wrote:

> Hi,
> I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 
> 
> You notein your docs that this is not a Tomcat issue per se, but JSSE issue.
> Please note that allowUnsafeLegacyRenegotiation is set to false.

Are you using the BIO connector?  This option is specific to that connector.  It would be helpful if you could share your connector configuration.

What version of the JVM are you using?  Can you upgrade to a version that resolves this issue?

Dan



> Looking into the source code I see the following:
> 
> public void handshake(Socket sock) throws IOException {
> 
>     ((SSLSocket)sock).startHandshake(); 
> 
>     if(!allowUnsafeLegacyRenegotiation) {
>        // disable all ciphers, avoiding any subsequent handshake
>        ((SSLSocket)sock).setEnabledCipherSuites(new String[0]);
>     }
> }
> 
> Also looking into Tomcat6/7 source code I see that the only difference is to check for existence of TLS_EMPTY_RENEGOTIATION_INFO_SCSV in JSSE. 
> 
> But other than that the logic is the same. So I can only assume that upgrading to Tomcat 6 would not solve my problem.
> Since I have verified via debugging also that allowUnsafeLegacyRenegotiation is indeed false, why do I get reports on SSL client renegotiation vulnerability? 
> 
> 
> I see in your notes (http://tomcat.apache.org/security-5.html) that:
> Requires JRE that supports RFC 5746. For Oracle JRE that is known to be 6u22 or later. 
> 
> 
> But as I understand the code, even if JRE does not support the RFC you still disable the handshake with your "hack" to set no enable ciphers
> 
> 
> Is this a Java/JSSE problem? What can I do? Can you please help me?
> 
> 
> Thank you!

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org

Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by Hermes Flying <fl...@yahoo.com>.
Hi, yes I am sorry for not mentioning but I am using BIO connector. 

I am using IBM JRE 6. 

Not Oracle/Sun. That is why I am trying to figure out what component (Tomcat/Java) is the one that has the issue.
I am not 100% it is the latest.




________________________________
 From: Daniel Mikusa <dm...@vmware.com>
To: Tomcat Users List <us...@tomcat.apache.org> 
Sent: Tuesday, November 27, 2012 12:09 AM
Subject: Re: Tomcat ssl vulnerability CVE-2009-3555
 
On Nov 26, 2012, at 3:36 PM, Hermes Flying wrote:

> Hi,
> I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 
> 
> You notein your docs that this is not a Tomcat issue per se, but JSSE issue.
> Please note that allowUnsafeLegacyRenegotiation is set to false.

Are you using the BIO connector?  This option is specific to that connector.  It would be helpful if you could share your connector configuration.

What version of the JVM are you using?  Can you upgrade to a version that resolves this issue?

Dan



> Looking into the source code I see the following:
> 
> public void handshake(Socket sock) throws IOException {
> 
>     ((SSLSocket)sock).startHandshake(); 
> 
>     if(!allowUnsafeLegacyRenegotiation) {
>        // disable all ciphers, avoiding any subsequent handshake
>        ((SSLSocket)sock).setEnabledCipherSuites(new String[0]);
>     }
> }
> 
> Also looking into Tomcat6/7 source code I see that the only difference is to check for existence of TLS_EMPTY_RENEGOTIATION_INFO_SCSV in JSSE. 
> 
> But other than that the logic is the same. So I can only assume that upgrading to Tomcat 6 would not solve my problem.
> Since I have verified via debugging also that allowUnsafeLegacyRenegotiation is indeed false, why do I get reports on SSL client renegotiation vulnerability? 
> 
> 
> I see in your notes (http://tomcat.apache.org/security-5.html) that:
> Requires JRE that supports RFC 5746. For Oracle JRE that is known to be 6u22 or later. 
> 
> 
> But as I understand the code, even if JRE does not support the RFC you still disable the handshake with your "hack" to set no enable ciphers
> 
> 
> Is this a Java/JSSE problem? What can I do? Can you please help me?
> 
> 
> Thank you!

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org

Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by Daniel Mikusa <dm...@vmware.com>.
On Nov 26, 2012, at 3:36 PM, Hermes Flying wrote:

> Hi,
> I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 
> 
> You notein your docs that this is not a Tomcat issue per se, but JSSE issue.
> Please note that allowUnsafeLegacyRenegotiation is set to false.

Are you using the BIO connector?  This option is specific to that connector.  It would be helpful if you could share your connector configuration.

What version of the JVM are you using?  Can you upgrade to a version that resolves this issue?

Dan



> Looking into the source code I see the following:
> 
> public void handshake(Socket sock) throws IOException {
> 
>     ((SSLSocket)sock).startHandshake(); 
> 
>     if(!allowUnsafeLegacyRenegotiation) {
>        // disable all ciphers, avoiding any subsequent handshake
>        ((SSLSocket)sock).setEnabledCipherSuites(new String[0]);
>     }
> }
> 
> Also looking into Tomcat6/7 source code I see that the only difference is to check for existence of TLS_EMPTY_RENEGOTIATION_INFO_SCSV in JSSE. 
> 
> But other than that the logic is the same. So I can only assume that upgrading to Tomcat 6 would not solve my problem.
> Since I have verified via debugging also that allowUnsafeLegacyRenegotiation is indeed false, why do I get reports on SSL client renegotiation vulnerability? 
> 
> 
> I see in your notes (http://tomcat.apache.org/security-5.html) that:
> Requires JRE that supports RFC 5746. For Oracle JRE that is known to be 6u22 or later. 
> 
> 
> But as I understand the code, even if JRE does not support the RFC you still disable the handshake with your "hack" to set no enable ciphers
> 
> 
> Is this a Java/JSSE problem? What can I do? Can you please help me?
> 
> 
> Thank you!

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by Christopher Schultz <ch...@christopherschultz.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hermes,

On 11/26/12 3:36 PM, Hermes Flying wrote:
> Hi, I am running Tomcat 5.35 and I got a report that it is
> vulnerable to SSL client renegotiation DoS.
> 
> You notein your docs that this is not a Tomcat issue per se, but
> JSSE issue. Please note that allowUnsafeLegacyRenegotiation is set
> to false. Looking into the source code I see the following:
> 
> public void handshake(Socket sock) throws IOException {
> 
> ((SSLSocket)sock).startHandshake();
> 
> if(!allowUnsafeLegacyRenegotiation) { // disable all ciphers,
> avoiding any subsequent handshake 
> ((SSLSocket)sock).setEnabledCipherSuites(new String[0]); } }
> 
> Also looking into Tomcat6/7 source code I see that the only
> difference is to check for existence of
> TLS_EMPTY_RENEGOTIATION_INFO_SCSV in JSSE.
> 
> But other than that the logic is the same. So I can only assume
> that upgrading to Tomcat 6 would not solve my problem. Since I have
> verified via debugging also that allowUnsafeLegacyRenegotiation is
> indeed false, why do I get reports on SSL client renegotiation
> vulnerability?

Many vulnerability scanners blindly check version numbers and then
just regurgitate information found in CVE reports. It's much easier
than actually probing for the actual vulnerability, and it generates
huge numbers of false positives.

It's very possible that your server is actually not vulnerable but
that you version number has been flagged as /potentially/ vulnerable.
What tool/service are you using to scan for vulnerabilities?

> I see in your notes (http://tomcat.apache.org/security-5.html)
> that: Requires JRE that supports RFC 5746. For Oracle JRE that is
> known to be 6u22 or later.

Allowing "safe" renegotiation requires this. Disabling unsafe
renegotiation does not require 6u22 or later -- because of the trick
of disabling all ciphers once the initial SSL handshake has been
completed.

> But as I understand the code, even if JRE does not support the RFC
> you still disable the handshake with your "hack" to set no enable
> ciphers

+1

> Is this a Java/JSSE problem? What can I do? Can you please help
> me?

I think you might want to put this into the "prove the vulnerability
exists" pile and go back to the tool/service for verification.

- -chris
-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2.0.17 (Darwin)
Comment: GPGTools - http://gpgtools.org
Comment: Using GnuPG with undefined - http://www.enigmail.net/

iEYEARECAAYFAlC1TLAACgkQ9CaO5/Lv0PB+vgCggfXX4KmAy+1NE+D2KFYMqDpo
g6cAoKuv7u1RMVcyAIOIWTyLiNO19l/A
=dUeN
-----END PGP SIGNATURE-----

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by Hermes Flying <fl...@yahoo.com>.
Hi, fair points.
I got this report from 2 separate sources. The one I know for a fact comes from Nessus. The second I don't know if it is also from Nessus or some other tool (I will try to find out and let you know).
I understand what you are saying about upgrade, and I will see if I can do that, but I believe that this will just gain time not solve the problem.
The affected code (which is part of my original mail from org.apache.tomcat.util.net.jsse.JSSESocketFactory) is the same in all later versions of Tomcat as well. 






________________________________
 From: André Warnier <aw...@ice-sa.com>
To: Tomcat Users List <us...@tomcat.apache.org> 
Sent: Monday, November 26, 2012 11:41 PM
Subject: Re: Tomcat ssl vulnerability CVE-2009-3555
 
Hermes Flying wrote:
> Just to be clear. When I say report,  I mean a report from a security penetration test suite which reports that the server allows renegotiation
> 
> 
> 
> 
> ________________________________
>  From: Hermes Flying <fl...@yahoo.com>
> To: "users@tomcat.apache.org" <us...@tomcat.apache.org> Sent: Monday, November 26, 2012 10:36 PM
> Subject: Tomcat ssl vulnerability CVE-2009-3555
>  Hi,
> I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 

Hi.

I believe that Tomcat 5.35 does not exist. You probably mean 5.5.35.

You may first want to have a look at this page : http://tomcat.apache.org/tomcat-55-eol.html

To comment on your request for help, and without getting into the technical details :

You do not specify which "security penetration test suite" was used to get this result. Such tools are known to generate false positives from time to time, and naming the tool may trigger someone's memory.

Tomcat is free software, developed, maintained and supported by volunteers.  As is human and logical, they like to dedicate more of their time to recent and current versions of Tomcat, rather than old ones, particularly after their end of life has been reached.
That may be considered as a reasonable trade-off for being able to use software that is free of charge.

To your own benefit thus : you would probably have a much better chance of getting attention and help for such an issue, if you installed a recent version of Tomcat, and confirmed with the same tool that you are getting the same result (or not)
(rather than "supposing" that you would.)
(have you tried to upgrade at least to v 5.5.36 (which is the "most current" release of that same branch), and checked if the same issue exists ?)

If you then do *not* see the same issue, there is a reasonable chance that the recommendation that will be made, is to upgrade Tomcat to this more recent version.
Or else, you will have to provide reasonable motives for which you cannot do that.

But if you *do* see the same issue with a very recent version, then it is almost guaranteed that you will get immediate attention.

All that does not mean that there will not be someone on this list that is willing to dedicate time to your issue, but you may be willing to increase your chances anyway.


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org

Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by André Warnier <aw...@ice-sa.com>.
Hermes Flying wrote:
> Just to be clear. When I say report,  I mean a report from a security penetration test suite which reports that the server allows renegotiation
> 
> 
> 
> 
> ________________________________
>  From: Hermes Flying <fl...@yahoo.com>
> To: "users@tomcat.apache.org" <us...@tomcat.apache.org> 
> Sent: Monday, November 26, 2012 10:36 PM
> Subject: Tomcat ssl vulnerability CVE-2009-3555
>  
> Hi,
> I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 
> 

Hi.

I believe that Tomcat 5.35 does not exist. You probably mean 5.5.35.

You may first want to have a look at this page : http://tomcat.apache.org/tomcat-55-eol.html

To comment on your request for help, and without getting into the technical details :

You do not specify which "security penetration test suite" was used to get this result. 
Such tools are known to generate false positives from time to time, and naming the tool 
may trigger someone's memory.

Tomcat is free software, developed, maintained and supported by volunteers.  As is human 
and logical, they like to dedicate more of their time to recent and current versions of 
Tomcat, rather than old ones, particularly after their end of life has been reached.
That may be considered as a reasonable trade-off for being able to use software that is 
free of charge.

To your own benefit thus : you would probably have a much better chance of getting 
attention and help for such an issue, if you installed a recent version of Tomcat, and 
confirmed with the same tool that you are getting the same result (or not)
(rather than "supposing" that you would.)
(have you tried to upgrade at least to v 5.5.36 (which is the "most current" release of 
that same branch), and checked if the same issue exists ?)

If you then do *not* see the same issue, there is a reasonable chance that the 
recommendation that will be made, is to upgrade Tomcat to this more recent version.
Or else, you will have to provide reasonable motives for which you cannot do that.

But if you *do* see the same issue with a very recent version, then it is almost 
guaranteed that you will get immediate attention.

All that does not mean that there will not be someone on this list that is willing to 
dedicate time to your issue, but you may be willing to increase your chances anyway.


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Re: Tomcat ssl vulnerability CVE-2009-3555

Posted by Hermes Flying <fl...@yahoo.com>.
Just to be clear. When I say report,  I mean a report from a security penetration test suite which reports that the server allows renegotiation




________________________________
 From: Hermes Flying <fl...@yahoo.com>
To: "users@tomcat.apache.org" <us...@tomcat.apache.org> 
Sent: Monday, November 26, 2012 10:36 PM
Subject: Tomcat ssl vulnerability CVE-2009-3555
 
Hi,
I am running Tomcat 5.35 and I got a report that it is vulnerable to SSL client renegotiation DoS. 

You notein your docs that this is not a Tomcat issue per se, but JSSE issue.
Please note that allowUnsafeLegacyRenegotiation is set to false. Looking into the source code I see the following:

public void handshake(Socket sock) throws IOException {

    ((SSLSocket)sock).startHandshake(); 

    if(!allowUnsafeLegacyRenegotiation) {
       // disable all ciphers, avoiding any subsequent handshake
       ((SSLSocket)sock).setEnabledCipherSuites(new String[0]);
    }
}

Also looking into Tomcat6/7 source code I see that the only difference is to check for existence of TLS_EMPTY_RENEGOTIATION_INFO_SCSV in JSSE. 

But other than that the logic is the same. So I can only assume that upgrading to Tomcat 6 would not solve my problem.
Since I have verified via debugging also that allowUnsafeLegacyRenegotiation is indeed false, why do I get reports on SSL client renegotiation vulnerability? 


I see in your notes (http://tomcat.apache.org/security-5.html) that:
Requires JRE that supports RFC 5746. For Oracle JRE that is known to be 6u22 or later. 


But as I understand the code, even if JRE does not support the RFC you still disable the handshake with your "hack" to set no enable ciphers


Is this a Java/JSSE problem? What can I do? Can you please help me?


Thank you!